> I've opened issue #919 [1] to track it. Thanks! This is definitly a critical regression, caused by my recent mkdir-with-children changes. It looks like I broke the HTML-form style of mkdir, by assuming that POST t=mkdir would always be done programmatically.
I'm not yet sure what the best fix is. I suspect that it will be to change web/unlink.py's POSTUnlinkedCreateDirectory to pay attention to the Content-Type header, and only complain about non-empty bodies when it is not that www-form-encoded type. And of course we need a test that constructs a form-based mkdir call and makes sure it works. I'll think about this further tonight. oops, -Brian _______________________________________________ tahoe-dev mailing list [email protected] http://allmydata.org/cgi-bin/mailman/listinfo/tahoe-dev
